Are large language models (LLMs) sentient?

1. When David Chalmers mean when he says “sentience” which does he mean?
   • affective consciousness
   • reasoning & agency
   • human-level intelligence
   • goal-directed behavior
   • phenomenological consciousness
2. What are some reasons that Chalmers presents that LLMs are sentient?
   • LLMs can say that they are sentient
   • LLMs seem sentient
   • Panpsychism – everything is sentient
   • LLMs have conversational ability
   • LLMs pass the Turing test
   • LLMs have evidence of domain-general intelligence
3. What are some reasons that Chalmers presents that LLMs are not sentient?
   • LLMs lack a body
   • LLMs are not biological systems
   • LLMs don’t have human-level reasoning
   • LLMs lack a world model
   • LLMs don’t have a unified agency

4. Explain why Chalmers thinks that “unified agency” is a “strongish” critique of LLMs.

5. Explain Chalmers’ conclusion about the current status of AI sentience.

6. Do you have any questions that you want to discuss in class?